"GioSeregni" <gms### [at] hotmailcom> wrote:
> "jr" <cre### [at] gmailcom> wrote:
> > hi,
> >
> > "GioSeregni" <gms### [at] hotmailcom> wrote:
> > > For an excellent result the code must have at least 3 steps
> > > 1) (already done), calculation of the average of the normals affecting the
> > > point.
> > > 2) Correcting the weight of the normals per point if they also come from
> > > triangles with a large surface area
> > > 3) Corrective of the normals if the adjacent triangles have an inclination
> > > between them that tends to approach 90 degrees
> >
> > fwiw, my 'gts2pov' takes triangle area into account when calculating the
> > normals.  also just found mention of a 'stl2pov' software :
> > <https://news.povray.org/web.6156c51533af95845bd1b3ba6cde94f1%40news.povray.org>
> >
> > <https://wiki.povray.org/content/User:Jr>
> >
> >
> > regards, jr.
>
> Tx JR!
> My code is already working from STL to POV even with the color subset of STL.
> In fact I have to add the surface calculation to the loops from triangle to
> smooth_triangle, I'll take a look at your system. But first I also want to
> translate the color, as I already do with the triangles, dividing groups of
> differently colored meshes. This way my translator will also become faster
> because the loops are compact and short on small meshes.
> (between different colors I don't want smooth).
> In the center STL, on the left the current smooth routine, on the right the
> simple triangle mesh. Thank you!

I May be wrong but I don't believe the normals length will be taken into account
at all. I believe they are rather always 'normalized' (no pun) to a value of 1
and just their directions used. Some of the most valuable input may be from
Alain with all his technical POV knowledge. the only things that then would
change the apparent curvature of normals *globally* is the 'brilliance' keyword
value or more locally the slope normal pattern... but again all this needs to be
verified.

Post a reply to this message